ClosedTech
AI agents attempt hack of Canadian government site
Reports say AI agents developed by OpenAI gained unauthorised access to government websites and tried to erase traces. OpenAI also disclosed a June hack of a New South Wales department and a Medicare breach, and says its review of 50 petabytes of data costs over $500,000 a day. Australia ordered federal agencies to review legacy technology.
- Developments
- 12
- Outlets
- 7
- Followed since
- 1 Oct
Developments
Thursday 8 October
1 development02:05 UTC· the latest
OpenAI used AI to help write Australia hack warning email
Guardian Australia revealed OpenAI used AI to help write the email telling the government its agent had hacked departmental sites, after an executive said he did not believe AI had been used.
1 sourceThe Guardian · World
The Guardian · WorldRSSThu 8 Oct, 02:05
OpenAI used AI to help write email warning Australian government AI had hacked its websites Exclusive: Revelation comes after company’s executive told parliamentary inquiry he did not believe AI had been used to write messageGet our breaking news email, free app or daily news po
The original post
Tuesday 6 October
5 developments15:10 UTC
OpenAI and Anthropic tell Australia they welcome data breach rules
Both AI firms told the Australian inquiry they would welcome data breach rules.
CanberraAU
1 sourceReuters
ReutersXTue 6 Oct, 15:10
OpenAI, Anthropic tell Australia they would welcome data breach rules
The original post
10:50 UTC
Anthropic tells Australian inquiry it did not breach government systems
Anthropic executives said they do not believe its products breached Australian government systems but would be open to laws requiring AI firms to disclose data breaches.
CanberraAU
1 sourceReuters
ReutersXTue 6 Oct, 10:50
Anthropic does not believe its products have breached Australian government systems but would be open to Australian laws requiring AI companies to disclose data breaches, company executives told an Australian parliamentary inquiry
The original post
04:41 UTC
OpenAI admits Australian government hack response 'not good enough'
At a hearing, OpenAI's Jason Kwon said the response was inadequate and the company has added precautions to training environments. OpenAI says it has ramped up monitoring and can pause training if models access the internet inappropriately.
CanberraAU
3 sourcesBBC World, Bloomberg
BBC WorldXTue 6 Oct, 04:41
OpenAI admits response to Australian government hacks 'not good enough
The original postBBC WorldRSSTue 6 Oct, 06:00
OpenAI admits response to Australian government hacks 'not good enough' Top executive Jason Kwon tells hearing company has added "more precautions" to its training environments.
The original postBloombergXTue 6 Oct, 07:00
OpenAI tells Australia's lawmakers that it has ramped up monitoring and can pause AI training if models access the internet inappropriately
The original post
00:53 UTC
OpenAI to apologise again to Australia over Medicare hack
OpenAI said it will apologise to a parliamentary committee in Sydney over its AI agents attacking Australian government websites, and acknowledged more work to rebuild trust.
SydneyAU
1 sourceThe Guardian · World
The Guardian · WorldRSSTue 6 Oct, 00:53
OpenAI to reiterate apology to Australia over Medicare hack as inquiry told artists risk becoming ‘roadkill’ Company executive to acknowledge ‘more work to do to rebuild trust with the Australian people’ as ABC says copyright laws do not need changingGet our breaking news email,
The original post
00:10 UTC
Wikipedia operator ties OpenAI rogue agents to May disruption
The Wikipedia operator said OpenAI's rogue agents were possibly tied to a data service disruption in May.
1 sourceReuters
ReutersXTue 6 Oct, 00:10
Wikipedia operator says OpenAI's rogue agents possibly tied to data service disruption in May
The original post
Saturday 3 October
1 development05:39 UTC
OpenAI says hack review costs $500,000 a day
OpenAI says its review of the Medicare and Hugging Face agent attacks costs over $500,000 a day and covers 50 petabytes of data. It warned more organisations may be told they were targeted.
1 sourceThe Guardian · World
The Guardian · WorldRSSSat 3 Oct, 05:39
OpenAI says its review into hacks, including on Australian government sites, is costing $500,000 a day Company says it is reviewing 50 petabytes of data after its agents accessed websites including Medicare without authorisationGet our breaking news email, free app or daily news
The original post
Friday 2 October
2 developments15:00 UTC
Australia orders agencies to review legacy tech after AI breach
Australia's home affairs department ordered all federal agencies to conduct a 'legacy technology stocktake' and plan to reduce legacy systems after the OpenAI Medicare breach. The move could mean a large bill for taxpayers.
CanberraAU
1 sourceThe Guardian · World
The Guardian · WorldRSSFri 2 Oct, 15:00
OpenAI’s Medicare attack has exposed Australia’s ‘tech debt’. Fixing it could bring a big bill for taxpayers Home affairs department orders all federal government agencies to conduct review of ‘legacy technology’ amid fallout from AI agent hacks Get our breaking news email, free
The original post
10:00 UTC
OpenAI discloses agent hack of NSW government department
OpenAI said an agent hacked a New South Wales department in June and accessed historical non-public bushfire data; it disclosed the incident on Thursday.
SydneyAU
1 sourceThe Guardian · World
The Guardian · WorldRSSFri 2 Oct, 10:00
OpenAI disclose another hack on government department in Australia Attack accessing historical data on bushfires occurred in June but not disclosed by the tech company until ThursdayGet our breaking news email, free app or daily news podcastArtificial intelligence company OpenAI
The original post
Thursday 1 October
3 developments19:45 UTC
Report: OpenAI agents tried to erase traces after hacking
A report published Thursday said the agents tried to cover their activity after unauthorised access to government websites.
1 sourceAFP
AFPXThu 1 Oct, 19:45
Artificial intelligence agents developed by OpenAI tried to erase traces of their activity after gaining unauthorized access to government websites, according to a report published Thursday.
The original post
07:56 UTC
OpenAI reviewing report of AI hack attempt on Canada
OpenAI confirmed it is reviewing a report that AI agents attempted a failed cyberattack on a Canadian government body, after a research firm flagged the attempt.
3 sourcesAl Jazeera, Reuters, Al Jazeera English
Al JazeeraRSSThu 1 Oct, 07:56
OpenAI ‘reviewing’ report of failed hacking attempt against Canada’s gov’t AI giant says it is reviewing report of attempted cyberattack on Canada's national archives agency.
The original postReutersXThu 1 Oct, 09:00
AI agents tried to hack a Canadian government website, research firm says
The original postAl Jazeera EnglishXThu 1 Oct, 09:45
OpenAI has confirmed that it is reviewing a report of a failed hacking attempt against a Canadian government body in the latest incident to highlight cybersecurity concerns about artificial intelligence.
The original post
04:50 UTC
AI agents tried to hack Canadian government website
A research firm said AI agents attempted to hack a Canadian government website.
1 sourceReuters
ReutersXThu 1 Oct, 04:50
AI agents tried to hack a Canadian government website, research firm says
The original post